92    QUANTITY AND QUALITY IN SOCIAL RĽSĽARCH
on case studies in much qualitative research detracts from the investigator's ability to generalize his or her findings has been critically addressed. In drawing attention to the importance of theoretical reasoning as the crux of the issue of case study generalization, the question of the relationship between theory and research again raises its head.
Notes
1 The injunction to empathize and 10 sec through the eyes of one's subjects can cause the qualitative researcher problems in certain domains. Billig (1977) has written about the difficulties associated with qualitative research on fascists; in particular, he argues that adopting a posture of disinterest or of sympathetic understanding is inappropriate in such research. Robbins. Anthony and Curtis (1973) have drawn attention to the difficulties of maintaining a stance of empathy without sympathy when studying religious groups like the 'Jesus Freaks', who define their view of the world as exclusively true.
2  Readers interested in (he considerable attention being given by anthropologists to the tenfativeness of their accounts of other cultures and to the devices they use to convey a sense of ethnographic authority should consult Clifford and Marcus (1986) and Marcus and Fischer (1986).
3  An example is Marcus's (1986) criticism of Willis's (1977) ethnographic research on working class lads. Marcus argues that Willis selected nonconformist boys and was uninterested in why some boys are nonconformist while others are not. Further, Marcus suggests that it is inappropriate to draw inferences about the operation of capitalism from such a sample,
5
The Debate about Quantitative and Qualitative Research
In this chapter, the main contrasting features of quantitative and qualitative research will be etched out. Much of the discussion in the literature on these two research traditions has created a somewhat exaggerated picture of their differences. These discussions reflect a tendency to treat quantitative and qualitative research as though they are mutually antagonistic ideal types of the research process. This tendency can be clearly discerned in some of the programmatic statements relating to qualitative research (e.g. J. Lofland, 1971; Bogdan and Taylor, 1975). While there are differences between the two research iraditions, as the first section of this chapter will explicate, there are also a number of points at which the differences are not as rigid as the programmatic statements often imply. Consequently, in addressing some of the contrasting features in quantitative and qualitative research, some areas of similarity will also be appraised. The discussion will then proceed to an assessment of the degree to which epistemological issues He at the heart of the contrast, or whether it is more a matter of different styles of data collection and analysis tout court. This issue has implications for the extent to which quantitative and qualitative research are deemed to be capable of integration (the focus of Chapter 6). It also has implications for the question of the extent to which quantitative and qualitative research constitute divergent models of (he research process, since it has been the suggestion that they represent distinct epistemologics that has played a major role in (he exaggeration of their differences. Finally, the question of whether these two research traditions share some common problems is examined.
Contrasting Features in Quantitative and Qualitative Research
Some of the main contrasting dimensions of quantitative and qualitative research have been either explicitly or implicitly explored in the previous chapters- This section will draw out these dif'
94    QUANTITY AND QUALITY IN SOCIAL RESEARCH
Tabic 5.1    Some differences between Quantitative and Qualitative Research
		Quantitative	Qualitative
(1)	Role of qualitative research	preparatory	means to exploration of actors' interpretations
(2)	Relationship between researcher and subject	distant	close
(3)	Researcher's stance in relation to subject	outsider	insider
(4)	Relationship between theory/concepts and research	confirmation	emergent
(5)	Research strategy	structured	unstructured
(6)	Scope of Undings	nomothetic	ideographic
(7)	Image of social reality	static and external to actor	processual and socially constructed by actor
(8)	Nature of data	hard, reliable	rich, deep
ferences more directly. Table 5.1 lists eight important dimensions on which the two research traditions diverge. The subsequent discussion explores these themes in some greater detail.
View of the Role of Qualitative Research
Quantitative researchers rarely totally deny the utility of qualitative research, but have tended to view it as an essentially exploratory way of conducting social investigations. Consequently, they have typically seen it as useful at the preparatory stage of a research project, a view which is clearly discernible in Blalock's (1970) attitude to participant observation which was quoted in Chapter 1. Precisely because of its exploratory and unstructured approach, qualitative research is often depicted as useful as a means of throwing up hunches and hypotheses which can be tested more rigorously by quantitative research. Such a view treats qualitative research as a somewhat second rate activity in implying that qualitative data
THE DEBÄTE    95
cannot stand in their own right because they need to be verified. The proponents of qualitative research see it as an end in itself, in particular because of its capacity to expose actors' meanings and interpretations, which is a central requirement of the approach and of its presumptive intellectual underpinnings which were discussed in Chapter 3. However, it is possible to detect a degree of unease among some qualitative researchers about the extent to which their findings can stand alone. Gans (1962, p. 350), at the end of his participant observation study of an Italian-American slum, exemplified this diffidence in proclaiming that his research 'is a reconnaissance - an initial exploration of a community to provide an overview', and went on to say: 'Many of the hypotheses reported here can eventually be tested against the results, of more systematic social science research.' Interestingly, in the second edition of the book which derived from this research, Gans.(1982, p. 414) has indicated that he would gat rid of this 'apologetic conclusion' if he were able to rewrite the book. He argues that the reason for this apologetic style was that at the time social scientists were strongly influenced by a belief in the appropriateness of the scientific method. By implication, Gans seems to be suggesting that the assertiveness of qualitative researchers coupled with the growing disillusionment with quantitative research have created a different climate, whereby investigations of the kind he undertook are increasingly regarded as ends in themselves.
Relationship be/ween Researcher and Subject In quantitative research the researcher's contact with the people being studied is fairly fleeting or even nonexistent. While the data collection phase often extends over many months, the contact with each individual is usually brief. In longitudinal surveys or in before-and-after experiments, the investigator returns to his or her subjects, but the degree of contact is still fairly short-lived. Indeed, the use of some methods associated with quantitative research may require no contact with subjects at all, except in an indirect sense; postal questionnaire surveys, laboratory experiments in which the researcher simply observes while others conduct the experiment (e.g. Milgram, 1963), and many forms of unobtrusive, structured observation (Webb et al., 1966) involve virtually no contact between researcher and subject. Even in interview surveys, the main investigator may have little or no contact with respondents since hired staff frequently carry out many (and sometimes all) of the interviews.
By contrast, qualitative research entails much more sustained contact, especially when participant observation is the central method. The degree to which there is sustained contact within a
96     QUANTITY AND QUALITY IN SOCIAL RESEARCH
particular study will vary a good deal; Gans (1962) had some con-lacl with KM) to 150 West Enders bul only twenty of Ihcse were intense The need for the fostering of such relationships is a produci o( the qualitative researcher's need lo sec the world through his or her subjects' eyes, since ihc researcher would be unable to gain any leverage on this level of analysis from a distance. Unstructured interviewing typically entails less sustained researcher-subject relationships than participant observation, but is invariably longer than survey interviews. In any case, the wide-ranging nature of the unstructured interview invariably necessitates a fairly close relationship between researcher and subject, which re-visits (which are relatively rare in survey interviewing) may intensify.
This contrast between the two research traditions can be illustrated through the work of Hirschi (1969) and Adler (1985). In the former case, self-administered questionnaires were the chief source of data. The questionnaires were administered io the children by their schools, so that Hirschi's contact with his subjects was minimal. By contrast, Adler had contact over a period of six years with some of the drug dealers she was investigating.
The Researcher's Stance in Relation to the Subject The ciuantitative researcher adopts the posture of an outsider looking in on the social world. He or she applies a pre-ordained framework on the subjects being investigated and is involved as little as possible in that world. This posture is the analogue of the detached scientific observer. Hirschi was chiefly concerned to test a theory of delinquency, and adopted a stance towards his subjects which entailed limited contact with them. They were merely fodder for the examination of his concerns, and not people with their own views and perspectives in relation to delinquent behaviour, school, and the other elements of Hirschi's research.
Among qualitative researchers there is a strong urge to'gel close'to the subjects being investigated - to be an insider. For qualitative researchers, it is only by getting close to their subjects and becoming an insider that they can view the world as a participant in that setting. Thus. Hirschi's outsider stance can be contrasted with Adlcr's view: 'the only way I could get close enough to [ upper-level drug dealers and smugglers] to discover what they were doing and to understand the world from their perspectives (Blumer, 1969) was to take a membership rolcin the setting' (Adler, 1985, p. 11). The insider standpoint may have its costs, the most frequently mentioned of which is the problem of 'going native', whereby the researcher loses his or her awareness of being a researcher and is seduced by the participants' perspective. Oakley (1984, p. 128), drawingon her research on becoming a mother, describes the experience of going native as follows:
THE DEBA IE    11
at three forty-five after two hours of a busy antenatal clinic 1 too would sigh with the doctors as we jointly peeped into the corridor and saw, still waiting, another row of abdomens ... Or at iwo in the morning I wanted someone io gel in there quickly and do a forceps delivery so 1 could (like them) go home to bed.
The experience of going native was not entirely negative in that it enabled her lo understand the pressures obstetricians are under. In any event, qualitative researchers are likely to see such drawbacks (if indeed they acknowledge them as such) as unavoidable consequences of a standpoint which is needed to gain access to their subjects' views. It is also apparent that the possibility of going native, with its implication of a loss of detachment, is to a significant degree incongruent with the image of impartial scientist which many quantitative researchers espouse.
Relationship between Theory/Concepts and Research The model of quantitative research presented in Figure 2.1 implies that theories and concepts arc the starting point for investigations carried out within its framework. Thus Hirschi in the Preface to Causes of Delinquency wrote: 'In this book I attempt to state and test a theory of delinquency.' By contrast, qualitative researchers often reject the idea of using theory as a precursor to an investigation (except perhaps as a means of providing an initial orientation to the situation as in 'grounded theory') since it may not reflect subjects' views about what is going on and what is important. Consequently, as one advocate of qualitative research has put it, 'It is marked by a concern with the discovery of theory rather than the verification of theory' (Filstcad, 1979, p. 38). Thus Adlcr's (1985) chief theoretical contribution - the notion of upper-level drug dealing as a component of a hedonistic life-style rather than an occupation - was an outcome of her research rather than a precursor to it.
In fact, the extent to which quantitative research is explicitly guided by theory has been questioned of many commentators. Instead, theoretical reasoning often occurs towards the end of the research process (Cicourel, 1982, pp. 19—20). Indeed, quantitative research is often much more exploratory and unpredictable in outcome than its descrip-lion by the advocates of qualitative research seems to imply. An example of the misleading nature of the view that quantitive research is devoid of surprise is the study of the International Typographical Union in the USA by Lipsct, Trow, and Coleman (1956), which involved a mixture of qualitative research and survey data. The latter were compiled in order to examine inter alia the relationship between union shop size and members' political involvement. However,
98     QUAM 11 >     IND  '.'IM' ITY IN SOCIAL  HI'SI AILCII
This analysis did ihm merely lesí hypotheses already held before the survey was conducted. Rather, the earlier hypotheses pointed to a fruitful line of enquiry, but many of the ideas and insights regarding the bearing of shop si/c on union polit ics emerged only in the course of the analysis of I he survey data. (I.ipsct, 1964, pp. 116-17)
Upset(1964, pp. MI-l2)alsoprovidesanumbcrofotherexamplcsof the way in which his survey data were a source of surprise. Similarly, Pugh (1988). writing about the Aston research which was discussed in Chapter 2, has commented on his disappointment that the effect of organization si/c on structure was so pervasive, an observation which can be interpreted as indicative of a certain element of surprise.
Quantitative research is often depicted as a routine practice whereby theories and their integral concepts are simply opcra-lionali/cd with a view to verifying their validity (see, for example, ľilsteaďs remark on this issue in the passage cited on p. 97). Ironically, some qualitative research is showing an explicit concern with theory, not solely as something which emerges from the data, but also as a phase in the research process which is formulated at the outset (Woods, 1986, pp. 156-61). Some of the school ethnographies cited in Chapters 3 and 4 show signs of a movement in this direction. Consequently, the contrast between quantitative and qualitative research in terms of verification of theory against preferring theory to emerge from the data is not as clear-cut as is sometimes implied.
Research Strategy
Quantitative research lends to adopt a structured approach to the study of society. To a large extent, this tendency is a product of the methods with which it is associated; both surveys and experiments require thai the issues to be focused upon be decided at the outset. In the previous section, the point was made that there is the possibility of an element of surprise in survey research which is frequently underestimated. However, it is evident that such investigations require that the variables be mapped out and introduced into the survey instruments. Survey research is structured in the sense that sampling and questionnaire construction are conducted prior to the start of data collection and then imposed on the sample members. Similarly, in experimental designs, independent and dependent variables, experimental and control groups are all part of the preparatory stajje. In both cases, once the research has been designed the broud shape of the findings can be discerned, that is. before one person has been interviewed or one experimental subject has received a treatment. An examination of lliischľs (1969) questionnaire reveals that he was fully aware of the material that needed
////    Kl IIAII      ')•>
io be collected in order to test the theories of delinquency which were to be examined.
By contrast, qualitative research lends io be more open. Many ethnographers advocate that I he delineation ol a research locus be deferred as long as possible (e.g. Cohen, 1978). Consequently, many qualitative researchers refer to a sensation of being overwhelmed during their early days in the field, since everything they observe is potentially 'data'. Whyte (1984), for example, sees ethnographic research as deriving much of its strength from its flexibility, which allows new leads to be followed up or additional data to be gathered in response to changes in ideas. But he also notes a limitation of such flexibility since 'you may find so many interesting things to study that you are at a loss to delimit the scope of your project and focus on specific problems' (Whyic, 1984, p. 225). Barrett (1976) has shown how his use of a prior theoretical framework (and a Structured research strategy which derived from this framework), which was used to guide his anthropological research on the factors associated with the economic success of a village in Nigeria, caused him to misconstrue his data. Initially, his theoretical focus had led him to believe that his data implied that the communal economic organization of the village, rather than religion, was the main contributor to its success. Almost two years after the completion of his thesis which was based on this field-work, Barrett felt impelled to accord religion a stronger role, following some critical comments he had received on his explanation. However, Barrett has since revised his explanation yet again to provide what he believes Io be the most accurate explanation of the village's success. This explanation differs radically from its predecessors in thai it vm»cs I.ii beyond the simple juxtaposition of religion and communal organization and emphasizes the emergence of development as a paramount goal within the village. Bui Barren's main point is that he was able to arrive at this more accurate account only when he had cast off the shackles of his prior theoretical framework. Such an experience underscores the strength of the qualitative researcher's preference for postponing theoretical reflection, albeit ai the possible cosl (noicd by Whyte) of being overwhelmed by data. It also suggests that il may be disadvantageous to ethnographers to structure their strategies in advance.
Also, the role of luck may be more apparent in such research, where being in the right place at the right time may significantly affect the direction of the research (Bryman, 1988), oi alternatively may give access to a potential research site (Buchanan, Boddy, and McCalman, 1988). For example, Bresnen (1988) relets to a lucky encounter in the pub soon after he had started his research on a
4